kernel 6.12.0-03-petasan missing modules
Pages: 1 2
admin
3,073 Posts
May 22, 2026, 8:06 pmQuote from admin on May 22, 2026, 8:06 pmThanks for the feedback.
Ok we will do extra tests since this is not the first report of issue.
Note that we already did a lot of tests to date for uograde to 4.1. And many users with no issues. But we will do further tests and will get back of we find anything.
Thanks for the feedback.
Ok we will do extra tests since this is not the first report of issue.
Note that we already did a lot of tests to date for uograde to 4.1. And many users with no issues. But we will do further tests and will get back of we find anything.
dbutti
41 Posts
May 23, 2026, 8:43 amQuote from dbutti on May 23, 2026, 8:43 amThank-you. I have no doubt that you perform a lot of tests for the new versions, but the point is that many Petasan users are running clusters that have been created years ago, and several updates have been installed over time. For that reason, the end result when we install 4.1.0 today could be different from what you get simply upgrading from a clean 4.0.0.
In the case discussed year, the issue is related mostly to kernel modules which are not loaded automatically, and this can be the result of the specific upgrade path (given that you are using a kernel which differs from the default one used in the Linux distribution taken as the base).
Things could be made easier if we had a clean procedure to completely reinstall a cluster node, keeping/reimporting the contents of its OSDs without causing a traffic storm with dozens terabytes of data being moved around. Is this something you can elaborate on? Is there any procedure that can be attempted for that?
Thank-you in advance,
Thank-you. I have no doubt that you perform a lot of tests for the new versions, but the point is that many Petasan users are running clusters that have been created years ago, and several updates have been installed over time. For that reason, the end result when we install 4.1.0 today could be different from what you get simply upgrading from a clean 4.0.0.
In the case discussed year, the issue is related mostly to kernel modules which are not loaded automatically, and this can be the result of the specific upgrade path (given that you are using a kernel which differs from the default one used in the Linux distribution taken as the base).
Things could be made easier if we had a clean procedure to completely reinstall a cluster node, keeping/reimporting the contents of its OSDs without causing a traffic storm with dozens terabytes of data being moved around. Is this something you can elaborate on? Is there any procedure that can be attempted for that?
Thank-you in advance,
Pages: 1 2
kernel 6.12.0-03-petasan missing modules
admin
3,073 Posts
Quote from admin on May 22, 2026, 8:06 pmThanks for the feedback.
Ok we will do extra tests since this is not the first report of issue.
Note that we already did a lot of tests to date for uograde to 4.1. And many users with no issues. But we will do further tests and will get back of we find anything.
Thanks for the feedback.
Ok we will do extra tests since this is not the first report of issue.
Note that we already did a lot of tests to date for uograde to 4.1. And many users with no issues. But we will do further tests and will get back of we find anything.
dbutti
41 Posts
Quote from dbutti on May 23, 2026, 8:43 amThank-you. I have no doubt that you perform a lot of tests for the new versions, but the point is that many Petasan users are running clusters that have been created years ago, and several updates have been installed over time. For that reason, the end result when we install 4.1.0 today could be different from what you get simply upgrading from a clean 4.0.0.
In the case discussed year, the issue is related mostly to kernel modules which are not loaded automatically, and this can be the result of the specific upgrade path (given that you are using a kernel which differs from the default one used in the Linux distribution taken as the base).
Things could be made easier if we had a clean procedure to completely reinstall a cluster node, keeping/reimporting the contents of its OSDs without causing a traffic storm with dozens terabytes of data being moved around. Is this something you can elaborate on? Is there any procedure that can be attempted for that?
Thank-you in advance,
Thank-you. I have no doubt that you perform a lot of tests for the new versions, but the point is that many Petasan users are running clusters that have been created years ago, and several updates have been installed over time. For that reason, the end result when we install 4.1.0 today could be different from what you get simply upgrading from a clean 4.0.0.
In the case discussed year, the issue is related mostly to kernel modules which are not loaded automatically, and this can be the result of the specific upgrade path (given that you are using a kernel which differs from the default one used in the Linux distribution taken as the base).
Things could be made easier if we had a clean procedure to completely reinstall a cluster node, keeping/reimporting the contents of its OSDs without causing a traffic storm with dozens terabytes of data being moved around. Is this something you can elaborate on? Is there any procedure that can be attempted for that?
Thank-you in advance,